Fuel ethanol is currently produced from feedstocks such as corn starch, sugar cane, and sugar beets. However, the potential for production of ethanol from these sources is limited as most of the farmland which is suitable for the production of these crops is already in use as a food source for humans. Furthermore, the production of ethanol from these feedstocks has a negative impact on the environment because fossil fuels used in the conversion process produce carbon dioxide and other byproducts.
The production of ethanol from cellulose-containing feedstocks, such as agricultural wastes, grasses, and forestry wastes, has received much attention in recent years. The reasons for this are because these feedstocks are widely available and inexpensive and their use for ethanol production provides an alternative to burning or landfilling lignocellulosic waste materials. Moreover, a byproduct of cellulose conversion, lignin, can be used as a fuel to power the process instead of fossil fuels. Several studies have concluded that, when the entire production and consumption cycle is taken into account, the use of ethanol produced from cellulose generates close to nil greenhouse gases.
The lignocellulosic feedstocks that are the most promising for ethanol production include (1) agricultural residues such as corn stover, wheat straw, barley straw, oat straw, rice straw, canola straw, and soybean stover; (2) grasses such as switch grass, miscanthus, cord grass, and reed canary grass; and (3) forestry wastes such as aspen wood and sawdust.
The first process step of converting lignocellulosic feedstock to ethanol involves breaking down the fibrous material to liberate sugar monomers, such as glucose, from the feedstock for conversion to ethanol in the subsequent step of fermentation. The two primary processes are acid hydrolysis, which involves the hydrolysis of the feedstock using a single step of acid treatment, and enzymatic hydrolysis, which involves an acid pretreatment followed by hydrolysis with cellulase enzymes.
In the acid hydrolysis process, the feedstock is subjected to steam and sulfuric acid at a temperature, acid concentration and length of time that are sufficient to hydrolyze the cellulose to glucose and the hemicellulose to xylose and arabinose. The sulfuric acid can be concentrated (25-80% w/w) or dilute (3-8% w/w). The glucose is then fermented to ethanol using yeast, and the ethanol is recovered and purified by distillation.
In the enzymatic hydrolysis process, the steam temperature, sulfuric acid concentration and treatment time are chosen to be milder than that in the acid hydrolysis process such that the cellulose surface area is greatly increased as the fibrous feedstock is converted to a muddy texture, but there is little conversion of the cellulose to glucose. The pretreated cellulose is then hydrolyzed to glucose in a subsequent step that uses cellulase enzymes, and the steam/acid treatment in this case is known as pretreatment. Prior to the addition of enzyme, the pH of the acidic feedstock is adjusted to a value that is suitable for the enzymatic hydrolysis reaction. Typically, this involves the addition of alkali to a pH of between about 4 to about 6, which is the optimal pH range for cellulases, although the pH can be higher if alkalophilic cellulases are used.
In one type of pretreatment process, the pressure produced by the steam is brought down rapidly with explosive decompression, which is known as steam explosion. Foody, (U.S. Pat. No. 4,461,648) describes the equipment and conditions used in steam explosion pretreatment. Steam explosion with sulfuric acid added at a pH of 0.4 to 2.0 has been the standard pretreatment process for two decades. It produces pretreated material that is uniform and requires less cellulase enzyme to hydrolyze cellulose than other pretreatment processes.
Cellulase enzymes catalyze the hydrolysis of the cellulose (β-1,4-D-glucan linkages) in the feedstock to products such as glucose, cellobiose, and other cellooligosaccharides. Cellulase is a generic term denoting a multienzyme mixture comprising exo-cellobiohydrolases (CBH), endoglucanases (EG) and β-glucosidases (βG) that can be produced by a number of plants and microorganisms. Cellulase enzymes work synergistically to hydrolyze cellulose to glucose. CBHI and CBHII generally act on the ends of the glucose polymers in cellulose microfibrils liberating cellobiose (Teeri and Koivula, Carbohydr. Europe, 1995, 12:28-33), while the endoglucanases act at random locations on the cellulose. Together, these enzymes hydrolyze cellulose to smaller cellooligosaccharides, primarily cellobiose. Cellobiose is hydrolyzed to glucose by β-glucosidase. It is known that most exo-cellobiohydrolases (CBH) and endoglucanases (EG) bind to cellulose in the feedstock via carbohydrate-binding modules (CBMs), such as cellulose-binding domains (CBDs), while most β-glucosidase enzymes, including Trichoderma and Aspergillus β-glucosidase enzymes, do not contain such binding modules and thus remain in solution. Cellulase enzymes may contain a linker region that connects the catalytic domain to the carbohydrate binding module. The linker region is believed to facilitate the activity of the catalytically active domain.
Cellulase enzymes containing a CBD have been produced by genetic engineering. For example, U.S. Pat. No. 5,763,254 (Wöldike et al.) discloses genetically engineered cellulose degrading enzymes derived from Humicola, Fusarium and Myceliopthora containing carbohydrate-binding domains. The goal of the studies was to produce cellulose or hemicellulose-degrading enzymes with novel combinations of the catalytically active domain, the linker region and the CBD or to produce CBD-containing cellulose or hemicellulose-degrading enzymes from those that lack a CBD. However, the ability of these novel enzymes to hydrolyze lignocellulosic feedstock was not demonstrated.
One significant problem with enzymatic hydrolysis processes is the large amount of cellulase enzyme required, which increases the cost of the process. The cost of cellulase accounts for more than 50% of the cost of hydrolysis. There are several factors that contribute to the enzyme requirement, but one of particular significance is the presence of compounds that reduce the reaction rate of cellulases and/or microorganisms in the subsequent fermentation of the sugar. For example, glucose released during the process inhibits cellulases, particularly β-glucosidase (Alfani et al., J. Membr. Sci., 1990, 52:339-350). Cellobiose produced during cellulose hydrolysis is a particularly potent inhibitor of cellulase (Tolan et al. in Biorefineries—Industrial Processes and Products, Vol. 1 Ed. Kamm et al., Chapter 9, page 203). Other soluble inhibitors are produced during pretreatment including sugar degradation products such as furfural and hydroxyl-methyl furfural; furan derivatives; organic acids, such as acetic acid; and soluble phenolic compounds derived from lignin. These compounds also inhibit yeast, which decreases ethanol production and consequently makes the process more costly. Although the effects of inhibitors can be reduced by performing the hydrolysis at a more dilute concentration, this requires the use of a large hydrolysis reactor, which adds to the expense of the process.
Simultaneous Saccharification and Fermentation (SSF) is a method of converting lignocellulosic biomass to ethanol which minimizes glucose inhibition of cellulases (see for example Ghosh et al., Enzyme Microb. Technol., 1982, 4:425-430). In an SSF system, enzymatic hydrolysis is carried out concurrently with yeast fermentation of glucose to ethanol. During SSF, the yeast removes glucose from the system by fermenting it to ethanol and this decreases inhibition of the cellulase. However, a disadvantage of this process is that the cellulase enzymes are inhibited by ethanol. In addition, SSF is typically carried out at temperatures of 35-38° C., which is lower than the 50° C. optimum for cellulase and higher than the 28° C. optimum for yeast. This intermediate temperature results in substandard performance by both the cellulase enzymes and the yeast. Thus, the inefficient hydrolysis requires very long reaction times and very large reaction vessels, both of which are costly.
Another approach that has been proposed to reduce inhibition by glucose, cellobiose, and other soluble inhibitors is removing hydrolysis products throughout hydrolysis by carrying out the reaction in a membrane reactor. A membrane reactor contains an ultrafiltration membrane which retains particles and high molecular weight components, such as enzyme, while allowing lower molecular weight molecules, such as sugars, to pass through the membrane as permeate.
An example of a process utilizing a membrane reactor is described in Ohlson and Trägårdh (Biotech. Bioeng., 1984, 26:647-653), in which the enzymatic hydrolysis of pretreated sallow (a willow tree species) is carried out in a reactor with a membrane having a 10,000 molecular weight cut off. Cellulases have a molecular weight of 50,000 and are therefore retained by the membrane in the hydrolysis reactor, while sugars are removed and replaced with buffer solution from a feed container with fresh substrate added intermittently. The rate of hydrolysis, as well as the yield of the soluble sugars, is enhanced due to the removal of inhibitors. However, a disadvantage of such reactors is that the membranes required for a commercial hydrolysis system are extremely large and expensive. The membranes are also prone to fouling by suspended solids present in the reaction mixture.
Various groups have investigated the recovery and recycling of cellulase enzymes during enzymatic hydrolysis to reduce the amount of the enzyme necessary during the conversion process. In some cases, this has also involved the continuous removal of hydrolyzates from the reaction mixture, thereby removing inhibitory compounds.
For example, Ishihara et al. (Biotech. Bioeng., 1991, 37:948-954) disclose the recycling of cellulase enzymes during the hydrolysis of steamed hardwood and hardwood kraft pulp in a reactor. The process involves the removal of a cellulase reaction mixture from the reactor, followed by the removal of insoluble residue containing lignin from the mixture by filtering with suction. The cellulase enzymes that are in the filtrate are separated from hydrolysis products, such as glucose and cellobiose, by ultrafiltration and then returned to the hydrolysis reactor. As stated by the investigators, a disadvantage of this system is that the extra step of solids removal would be impractical in an industrial application due to the rise in the cost of raw material. In addition, most of the cellulases remain bound to the cellulose and are difficult to recover.
Larry et al. (Appl. Microbiol. Biotechnol., 1986, 25:256-261) describe an approach for the re-use of cellulases which involves performing the hydrolysis in a column reactor containing cellulose (Solka Floc). The hydrolyzed sugars are continuously removed by percolating the column with a steady stream of buffer. According to the investigators, the removal of sugar products should reduce product inhibition and enhance hydrolysis efficiencies. However, inadequate hydrolysis is obtained since unbound β-glucosidase and endoglucanase elute from the column.
Knutsen and Davis (Appl. Biochem. Biotech., 2002, 98-100:1161-1172) report a combined inclined sedimentation and ultrafiltration process for recovering cellulase enzymes during the hydrolysis of lignocellulosic biomass. The goal of the process is to remove larger lignocellulosic particles so a membrane filter does not become clogged during a subsequent step of ultrafiltration. The process first involves treating lignocellulosic particles with cellulase enzymes and then feeding the resulting mixture into an inclined settler. Large lignocellulosic particles, including enzyme bound to the particles, are retained in the inclined settler, while smaller particles and soluble enzyme are carried out with the settler overflow. The overflow is then fed to a crossflow ultrafiltration unit to recover unbound cellulases, while allowing for the passage of sugars. After ultrafiltration, the recovered cellulases are added to the hydrolysis reactor. The lignocellulosic particles remaining in the inclined settler, along with the bound enzyme, are returned to the reactor along with the settler underflow. One disadvantage of this system is that the operation of such a system on the scale of a commercial hydrolysis reactor, which is likely to be about 70 feet tall and process thousands of gallons of slurry every hour, will be prohibitively difficult. A second disadvantage of this system is that the concentration of glucose and cellobiose in the reactor remains unchanged throughout the process so that a high level of inhibition still occurs. A further disadvantage of the process is that it requires an expensive ultrafiltration step to recover unbound cellulases.
Mores et al. (Appl. Biochem. Biotech., 2001, 91-93:297-309) report a combined inclined sedimentation and ultrafiltration process similar to that described by Knutsen and Davis (supra). However, the process of Mores et al. involves an extra clarification step involving subjecting the settler overflow to microfiltration prior to ultrafiltration to reduce fouling of the ultrafiltration membrane. The process of Mores et al. would be subject to the same limitations as those described for Knutsen and Davis (supra).
U.S. Pat. No. 3,972,775 (Wilke et al.) reports a process for recycling cellulase in which the hydrolysis products are separated into an aqueous sugar-containing phase and a solid phase containing unhydrolyzed spent solids after the hydrolysis is complete. The spent solids are washed with water to recover enzyme adsorbed on it and the resulting wash water containing the desorbed enzyme is fed to the hydrolysis reaction. The remaining spent solids can be used as a source of fuel for the system. However, the process of Wilke et al. incurs the cost of the additional water wash after the hydrolysis, which is significant due to the large amount of solid material and the fine particulate nature of the solids. In addition, the process does not result in the removal of inhibitors of cellulase enzymes present during the hydrolysis reaction since the separation of hydrolyzates is carried out after completion of the hydrolysis reaction.
Ramos et al. (Enzyme Microb. Technol., 1993, 15:19-25) disclose a process in which steam-exploded eucalyptus chips are hydrolyzed using cellulase with removal of soluble sugars and the recycling of enzyme. The process involves stopping the reaction at selected incubation times and collecting the unhydrolyzed, enzyme-containing residue on a sintered glass filter. The enzyme-containing residue is washed with hydrolysis buffer to remove soluble sugars. The washed residue is then re-suspended in fresh hydrolysis buffer containing fresh β-glucosidase enzyme and incubated at 45° C. for subsequent hydrolysis. A problem with this process is that the repeated addition of fresh β-glucosidase after re-suspension would significantly increase the expense of the process.
Lee et al. (Biotech. Bioeng., 1994, 45:328-336) examine the recycling of cellulase enzymes in a procedure involving over five successive rounds of hydrolysis. The process involves adding cellulase enzymes and β-glucosidase (Novozym® 188) to peroxide-treated birch and recovering the residual substrate by filtering after 12 hours of hydrolysis. Fresh substrate is then added to the recovered residual substrate to achieve a total substrate concentration of 2% and the resulting mixture is re-suspended in buffer containing β-glucosidase and the hydrolysis is allowed to continue. Cellulase recycling followed by hydrolysis is subsequently repeated three times. Also disclosed is a procedure for recycling cellulases present in the complete reaction mixture both before and after all the cellulose is hydrolyzed. Similar to Ramos et al., a limitation of this process is that β-glucosidase must be added to the reaction at each recycling step.
U.S. Pat. No. 5,962,289 (Kilbum et al.) discloses a three-step enzymatic hydrolysis. The first step of the process involves adding both endoglucanase and exoglucanase to a lignocellulosic material to be hydrolyzed to cellobiose. The second step involves adding this material to an Avicel® column to adsorb the endoglucanase and exoglucanase. In a third step, the eluent containing cellobiose is then applied to a second Avicel® column containing β-glucosidase immobilized via a CBD. The immobilized β-glucosidase hydrolyzes the cellobiose into glucose. One limitation of this method is that the production of glucose is carried out in three distinct process steps, which is highly complex and costly. A second limitation is that sending the slurry of partially-hydrolyzed lignocellulosic material through the column of Avicel® at a high flow rate typical of a commercial hydrolysis process is very difficult. In addition, the highly inhibitory effects of cellobiose are present during the cellulose hydrolysis.
At present, there is much difficulty in the art to operate an efficient enzymatic hydrolysis of cellulose. A key obstacle is overcoming the inhibitory effects of glucose and especially cellobiose to cellulase. The development of such a system remains a critical requirement for a process to convert cellulose to glucose.